Dynamic Range and Mass Accuracy of Wide-Scan ... - ACS Publications

School of Biosciences, and Department of Electrical, Electronic and Computer .... Analytical Chemistry 2012 84 (9), 3998-4005 .... Rapid Communication...
0 downloads 0 Views 277KB Size
Anal. Chem. 2007, 79, 4595-4602

Dynamic Range and Mass Accuracy of Wide-Scan Direct Infusion Nanoelectrospray Fourier Transform Ion Cyclotron Resonance Mass Spectrometry-Based Metabolomics Increased by the Spectral Stitching Method Andrew D. Southam,† Tristan G. Payne,‡ Helen J. Cooper,† Theodoros N. Arvanitis,‡ and Mark R. Viant*,†

School of Biosciences, and Department of Electrical, Electronic and Computer Engineering, School of Engineering, University of Birmingham, Edgbaston, Birmingham, B15 2TT, UK

Direct infusion nanoelectrospray Fourier transform ion cyclotron resonance mass spectrometry (DI nESI FT-ICR MS) offers high mass accuracy and resolution for analyzing complex metabolite mixtures. High dynamic range across a wide mass range, however, can only be achieved at the expense of mass accuracy, since the large numbers of ions entering the ICR detector induce adverse spacecharge effects. Here we report an optimized strategy for wide-scan DI nESI FT-ICR MS that increases dynamic range but maintains high mass accuracy. It comprises the collection of multiple adjacent selected ion monitoring (SIM) windows that are stitched together using novel algorithms. The final SIM-stitching method, derived from several optimization experiments, comprises 21 adjoining SIM windows each of width m/z 30 (from m/z 70 to 500; adjacent windows overlap by m/z 10) with an automated gain control (AGC) target of 1 × 105 charges. SIMstitching and wide-scan range (WSR; Thermo Electron) were compared using a defined standard to assess mass accuracy and a liver extract to assess peak count and dynamic range. SIM-stitching decreased the maximum mass error by 1.3- and 4.3-fold, and increased the peak count by 5.3- and 1.8-fold, versus WSR (AGC targets of 1 × 105 and 5 × 105, respectively). SIM-stitching achieved an rms mass error of 0.18 ppm and detected over 3000 peaks in liver extract. This novel approach increases metabolome coverage, has very high mass accuracy, and at 5.5 min/sample is conducive for high-throughput metabolomics. Metabolomics involves the measurement of multiple small molecules within a biological sample to generate a unique metabolic profile. These profiles can be compared directly between different biological phenotypes, and as metabolites represent the * Corresponding author. Phone: +44-(0)121-414-2219. Fax: +44-(0)121-4145925. E-mail: [email protected]. † School of Biosciences. ‡ Department of Electrical, Electronic and Computer Engineering, School of Engineering. 10.1021/ac062446p CCC: $37.00 Published on Web 05/19/2007

© 2007 American Chemical Society

end products of complex cellular control systems, such analyses can give insight into upstream gene and protein activity.1 This approach has been applied to many fields, notably toxicology, drug design, disease diagnosis, and quality control of food substances.2-6 Reproducibility of the measurement system is critical, with nuclear magnetic resonance (NMR)-based methods proving very robust.6,7 However, due to its superior sensitivity over NMR, mass spectrometry (MS) represents an attractive detection method for metabolomics.8,9 Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR MS) is a particularly powerful tool for complex mixture analysis due to its ultrahigh mass resolution and mass accuracy10,11 and has been applied successfully to a small number of metabolomics studies.12-16 In principle, this analysis method enables the empirical formulas of many low molecular weight metabolites to be unambiguously identified based upon (1) Fiehn, O.; Kopka, J.; Dormann, P.; Altmann, T.; Trethewey, R. N.; Willmitzer, L. Nat. Biotechnol. 2000, 18, 1157-1161. (2) Griffin, J. L.; Shockcor, J. P. Nat. Rev. Cancer 2004, 4, 551-561. (3) Nicholson, J. K.; Connelly, J.; Lindon, J. C.; Holmes, E. Nat. Rev. Drug Discovery 2002, 1, 153-161. (4) Rischer, H.; Oksman-Caldentey, K. M. Trends Biotechnol. 2006, 24, 102104. (5) Stoughton, R. B.; Friend, S. H. Nat. Rev. Drug Discovery 2005, 4, 345-350. (6) Viant, M. R.; Rosenblum, E. S.; Tjeerdema, R. S. Environ. Sci. Technol. 2003, 37, 4982-4989. (7) Nicholson, J. K.; Lindon, J. C.; Holmes, E. Xenobiotica 1999, 29, 11811189. (8) Dunn, W. B.; Ellis, D. I. Trends Anal. Chem. 2005, 24, 285-294. (9) Villas-Boas, S. G.; Mas, S.; Akesson, M.; Smedsgaard, J.; Nielsen, J. Mass Spectrom. Rev. 2005, 24, 613-646. (10) Hendrickson, C. L.; Emmett, M. R. Annu. Rev. Phys. Chem. 1999, 50, 517536. (11) Marshall, A. G.; Hendrickson, C. L.; Jackson, G. S. Mass Spectrom. Rev. 1998, 17, 1-35. (12) Cooper, H. J.; Marshall, A. G. J. Agric. Food Chem. 2001, 49, 5710-5718. (13) Fard, A. M.; Turner, A. G.; Willett, G. D. Aust. J. Chem. 2003, 56, 499508. (14) Stentiford, G. D.; Viant, M. R.; Ward, D. G.; Johnson, P. J.; Martin, A.; Wei, W. B.; Cooper, H. J.; Lyons, B. P.; Feist, S. W. Omics: J. Integrative Biol. 2005, 9, 281-299. (15) Aharoni, A.; Ric de Vos, C. H.; Verhoeven, H. A.; Maliepaard, C. A.; Kruppa, G.; Bino, R.; Goodenowe, D. B. Omics: J. Integrative Biol. 2002, 6, 217234. (16) Brown, S. C.; Kruppa, G.; Dasseux, J. L. Mass Spectrom. Rev. 2005, 24, 223-231.

Analytical Chemistry, Vol. 79, No. 12, June 15, 2007 4595

mass measurements alone. Furthermore, if isotope information is used in conjunction with accurate mass measurements, identification of peaks over a wide mass range can be achieved with even greater certainty.17 This is of considerable importance in metabolomics, as the identification of unknown metabolites remains one of the biggest analytical challenges. Liquid chromatography MS (LC-MS)18-20 and direct infusion electrospray MS (DI ESI MS)21-24 are both commonly used in metabolomics. One major benefit of DI ESI MS is that the dependent axis corresponding to the mass-to-charge ratio (m/z) is substantially more reproducible than in LC-MS, where chromatographic retention time (RT) will change as a function of column age and performance; e.g., a routine 1 ppm uncertainty in m/z in FT-ICR MS is equivalent to a 0.0018-s uncertainty in RT for a typical 30-min chromatographic separation, a value that is currently unachievable. The reproducibility of the m/z-dependent axis value is important for spectral interpretation as well as for facilitating comparison of data sets collected during different experimental sessions. Although DI ESI MS benefits from a fundamentally more simple one-dimensional data set than the twodimensional LC-MS data set, the availability of the time dimension in LC-MS can provide better discrimination of metabolites and aid in structural identification. Considering sample throughput, DI ESI MS offers shorter analysis times than conventional LC-MS, which is clearly an advantage. Conventional DI ESI MS, however, does suffer from ionization suppression and enhancement, which arises when particular analytes preferentially ionize over less polar metabolites in the complex mixture. The conventional method for reducing this problem is to utilize LC separation prior to MS analysis,25 which removes salts from the complex mixture and reduces the number of compounds entering the mass spectrometer at any given time. However, by slowing the sample delivery rate to 200 nL/min using a chip-based DI nanoelectrospray source (NanoMate, Advion Biosciences), as reported here, ionization suppression has been shown to be substantially reduced compared to conventional DI electrospray ionization.26 For samples analyzed by direct infusion, the finite dynamic range of a MS detector will limit the observation to only the most intense ions. To identify lower abundance ions, the dynamic range, i.e., the ratio of the highest to lowest concentration metabolites detected, needs to be increased. High dynamic range can be achieved using DI ESI FT-ICR MS by increasing the number of ions that enter the ICR detector cell. This, however, can signifi(17) Kind, T.; Fiehn, O. BMC Bioinformatics 2006, 7. (18) Idborg-Bjorkman, H.; Edlund, P. O.; Kvalheim, O. M.; Schuppe-Koistinen, I.; Jacobsson, S. P. Anal. Chem. 2003, 75, 4784-4792. (19) Plumb, R. S.; Stumpf, C. L.; Gorenstein, M. V.; Castro-Perez, J. M.; Dear, G. J.; Anthony, M.; Sweatman, B. C.; Connor, S. C.; Haselden, J. N. Rapid Commun. Mass Spectrom. 2002, 16, 1991-1996. (20) von Roepenack-Lahaye, E.; Degenkolb, T.; Zerjeski, M.; Franz, M.; Roth, U.; Wessjohann, L.; Schmidt, J.; Scheel, D.; Clemens, S. Plant Physiol. 2004, 134, 548-559. (21) Allen, J.; Davey, H. M.; Broadhurst, D.; Heald, J. K.; Rowland, J. J.; Oliver, S. G.; Kell, D. B. Nat. Biotechnol. 2003, 21, 692-696. (22) Castrillo, J. I.; Hayes, A.; Mohammed, S.; Gaskell, S. J.; Oliver, S. G. Phytochemistry 2003, 62, 929-937. (23) Dunn, W. B.; Overy, S.; Quick, W. P. Metabolomics 2005, 1, 137-148. (24) Goodacre, R.; York, E. V.; Heald, J. K.; Scott, I. M. Phytochemistry 2003, 62, 859-863. (25) Siuzdak, G. The Expanding Role of Mass Spectrometry in Biotechnology.; MCC Press: San Diego, CA., 2003. (26) Hop, C.; Chen, Y.; Yu, L. J. Rapid Commun. Mass Spectrom. 2005, 19, 31393142.

4596

Analytical Chemistry, Vol. 79, No. 12, June 15, 2007

cantly increase space-charge effects,27,28 which degrade the mass accuracy to such an extent that derivation of unique empirical formulas for the observed peaks is no longer possible. In this paper, we report a strategy for wide-scan DI nESI FT-ICR MS that effectively increases the overall dynamic range of the mass spectrum, enabling observation of both low- and high-concentration biological metabolites with high mass accuracy. The approach is based upon the collection of multiple narrow, overlapping spectra (or “windows”), that are subsequently combined (or “stitched”) together using novel algorithms. A somewhat similar non-data-dependent “wide-scan” acquisition strategy was reported by Venable et al.,29 based upon the sequential isolation and fragmentation of multiple m/z 10 wide precursor windows using LC-MS/MS, which increased the signal-to-noise ratio in a proteomics study. In our approach, each window is analyzed using selected ion monitoring (SIM) mode, and consequently, each window is composed of relatively few ions to minimize spacecharge effects. This approach produces a “SIM-stitched”, widescan mass spectrum with a significantly greater dynamic range, enabling many more metabolites to be detected, and with higher mass accuracy than is possible with existing methods. To maintain high mass resolution, relatively slow data acquisition is required (typically 1 s/transient) with many transients averaged to produce a single mass spectrum. Therefore, our method requires continuous infusion of each sample for a few minutes, which is achieved using a NanoMate chip-based nanoelectrospray system.30 This system has previously been used for metabolomics31 and is fully automated, using a new sample delivery tip and electrospray nozzle for each analysis to enable high throughput with no crosscontamination. Stitching the multiple SIM windows together to produce a single contiguous, wide-scan spectrum is important for several reasons. First, the multivariate analysis of mass spectral fingerprints using widely utilized methods in metabolomics, such as principal components analysis, requires a single spectral fingerprint per biological sample. Second, methods that search for characteristic peak spacings across the entire spectral range (for example, Breitling et al.32) also require one contiguous data set. Third, the stitching algorithm is used to mass calibrate SIM windows that do not contain internal calibrants (see Supporting Information for details). Finally, stitching multiple data sets together substantially aids data handling, storage, and visual inspection of the mass spectral measurements. The aims of this study are therefore to develop and test a novel analytical and bioinformatic approach for wide-scan, high dynamic range DI nESI FT-ICR MS analysis of complex biological mixtures of low molecular weight metabolites, which can achieve a maximum mass error of